Robust and accurate prediction of residue–residue interactions across protein interfaces using evolutionary information

By Sergey Ovchinnikov, Hetunandan Kamisetty, David Baker. 2014

601 citations recorded by the source corpus.

Do the amino acid sequence identities of residues that make contact across protein interfaces covary during evolution? If so, such covariance could be used to predict contacts across interfaces and assemble models of biological complexes. We find that residue pairs identified using a pseudo-likelihood-based method to covary across protein–protein interfaces in the 50S ribosomal unit and 28 additional bacterial protein complexes with known structure are almost always in contact in the complex, provided that the number of aligned sequences is greater than the average length of the two proteins. We use this method to make subunit contact predictions for an additional 36 protein complexes with unknown structures, and present models based on these predictions for the tripartite ATP-independent periplasmic (TRAP) transporter, the tripartite efflux system, the pyruvate formate lyase-activating enzyme complex, and the methionine ABC transporter. DOI: http://dx.doi.org/10.7554/eLife.02030.001

Subfields: Genetics, Molecular Biology, Computational Biology & Bioinformatics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Genetics

In Molecular Biology

In Computational Biology & Bioinformatics

Citations in the corpus

High-resolution de novo structure prediction from primary sequence (2022, neutral): “Gi,j = max  Fi−1,j, i > 0, j ≥ 0; Gi−1,j, i > 0, j ≥ 0; 0, i = 0, j = 0; −∞ (6)”
Evolutionary context-integrated deep sequence modeling for protein engineering (2020, background): “As structurally proximal protein residues are often inferred from co-variation pairs from sequence evolution historically [35,36] , we hypothesized that co-evolution information can also be used to infer epistasis or fitness of proteins.”
Coevolutionary Landscape Inference and the Context-Dependence of Mutations in Beta-Lactamase TEM-1 (2015, background): “[46] Sergey Ovchinnikov, Hetunandan Kamisetty, and David Baker.”
MetaPSICOV: combining coevolution methods for accurate prediction of contacts and long range hydrogen bonding in proteins (2014, background): “Contact: d.t.jones@ucl.ac.uk Supplementary information: Supplementary data are available at Bioinformatics online.”
Enhancing Evolutionary Couplings with Deep Convolutional Neural Networks (2017, methodology): “…applied to protein complex assembly and interactions, disordered region structure prediction, RNA structure prediction, and mutagenesis analysis (Ovchinnikov et al., 2014; Uguzzoni et al., 2017; Hopf et al., 2014, 2017; Toth-Petroczy et al., 2016; De Leonardis et al., 2015; Weinreb et al.,…”
Sequence co-evolution gives 3D contacts and structures of protein complexes (2014, background): “the same time as specific residue pair couplings across the proteins (as also shown in the work by the Baker lab on ribosomal protein interactions (Ovchinnikov et al., 2014)).”
Masked Language Modeling for Proteins via Linearly Scalable Long-Context Transformers (2020, neutral): “This follows recent works employing simpler statistical models that predict protein quaternary structure, protein-protein interactions and protein interaction networks from evolutionary sequence data [52, 26, 36, 3, 14].”
Large-scale determination of previously unsolved protein structures using evolutionary information (2015, methodology): “…from such predictions have been used to model a wide range of unknown protein structures ( Hayat et al., 2014 ; Wickles et al., 2014 ; Abriata, 2015 ; Antala et al., 2015 ; Hopf et al., 2015 ; Tian et al., 2015 ) and protein–protein complexes ( Ovchinnikov et al., 2014 ; Hopf et al., 2014 ).”
How Pairwise Coevolutionary Models Capture the Collective Residue Variability in Proteins? (2018, methodology): “Methods like the direct coupling analysis (DCA) [8, 9], PsiCov [10] or Gremlin [11] allow for the prediction of residue-residue contacts using sequence information alone, and can be used to predict three-dimensional protein structures [12, 13] and to assemble protein complexes [14, 15, 16].”
Inferring interaction partners from protein sequences using mutual information (2018, methodology): “To demonstrate the generality of the MI-IPA, we applied it to several pairs of protein families [13].”
Protein complex prediction with AlphaFold-Multimer (2021, background): “Such an ability has also been shown to be a useful factor for predicting protein-protein interactions [25, 26].”